| |
Time to Translate New Knowledge into Practice: A Call for a National Genital Herpes Control Program
|
| |
| |
EDITORIAL
The Journal of Infectious Diseases July 1, 2006;194:6-7
Edward W. Hook1,2 and Peter Leone3,4
1University of Alabama at Birmingham and 2Jefferson County Department of Health, Birmingham; 3University of North Carolina at Chapel Hill, Chapel Hill, and 4North Carolina State Department of Public Health, Raleigh
(See the article by Wald et al., below.)
The time is right for a national effort to control genital herpes, one of the most common sexually transmitted diseases (STDs) in the United States. With an estimated 45 million Americans infected, genital herpes causes considerable morbidity for those with infection, is an important source of congenital infections, and increases the risk of acquisition of HIV infection [1-3]. Genital herpes is also one of the few common STDs for which, at present, there is little coordinated emphasis on control efforts beyond a modest amount of ongoing research. The absence of a federally led control program is related, in part, to the fact most of the epidemic is unseen; about 90% of persons with this chronic viral infection are unaware that they are infected [1]. As a result, and because nearly all infected persons regularly shed the virus asymptomatically, sexual transmission of genital herpes continues. Without knowledge of infection on the part of most persons with genital herpes, there has been little public call for a broad-based campaign for genital herpes control.
Effective tools for prevention are required elements for any successful STD control effort. Although we now have those tools, until recently there have been many reasonable explanations for genital herpes not to have been highly prioritized for control efforts. Most importantly, despite the availability of several theoretically useful interventions, there were no proven measures for the prevention of spread: data demonstrating that condoms prevented transmission were lacking, and there were reasons to hypothesize why condoms might not be effective; in addition, although antiviral therapy had been shown to reduce asymptomatic viral shedding by infected persons, whether this effect resulted in reduced transmission was unknown; finally, although accurate serological tests had become available, there was no evidence that, even if infected persons were aware of their infection, they would take the necessary steps to reduce transmission to others. During the past 3 years, however, each of these knowledge deficits has been addressed.
The first concern to be successfully addressed was related to suppressive antiviral therapy. In 2004, results of a large, multicenter, randomized controlled trial were published that demonstrated not only that daily suppressive therapy significantly reduced the frequency and intensity of symptomatic and subclinical genital herpes recurrences but that this reduction in asymptomatic shedding translated into an about 50% reduction in transmission to sex partners [4]. Thus, chronic suppressive antiviral therapy is now a proven method for reducing genital herpes transmission. Even though this is important, because most persons with infection do not know that they are infected and because of cost considerations related to the purchase of continuing suppressive therapy, these data were not sufficient to generate enthusiasm for embarking on major control efforts.
So what about prevention measures for the about 90% of infected persons who are unaware of their infections? Condoms may be part of the answer. Recently, several publications have shown that condoms reduce the transmission of genital herpes [5, 6], overcoming the theoretical concern that, because condoms do not cover all potential sites of infection, they may not effectively reduce the transmission of infection. These data add to the armamentarium of proven tools available for genital herpes control efforts. However, given that condom promotion has been ongoing for the prevention of other STDs, including HIV, and the fact that >2 decades of efforts to promote condom use have been only partially successful, condom promotion may not be sufficient as a basis for a genital herpes prevention campaign either, although it is certainly a useful element of control strategies.
Knowledge of infection status is a necessary first step to reduce the transmission of genital herpes by those who do not know that they are infected. There are now type-specific serological assays that reliably distinguish between herpes simplex virus type 1 (HSV-1) and HSV-2 infections [7]. Although no serological assay is perfect, the performance of the current generation of type-specific serological assays for HSV infections are more than adequate for use in most clinical settings [8]. These are the tools that could be used to identify the large numbers of infected persons who are unaware that they have genital herpes, and serological test results can be provided to individuals without undue psychological distress [9]. But can knowledge of infection lead to changes in behavior that reduce the transmission of genital herpes?
The answer appears to be "yes." In this issue of the Journal of Infectious Diseases, Wald et al. [10] provide data that suggest that serological testing, coupled with efforts to encourage the disclosure of infection to sex partners, may provide a basis for broadly applicable efforts to control genital herpes. In their retrospective analysis of persons presenting with recently acquired genital herpes, Wald et al. found that the time to acquisition of infection was significantly increased in relationships in which the infected persons had disclosed to sex partners that they had genital herpes. These findings are contrary to the considerable skepticism that has been voiced regarding whether knowledge of infection might lead to increased disclosure or whether disclosure would prevent transmission. Although this study provides data only for those in whom transmission did occur (and thus are biased toward the null because they provide a measurement only of those in whom efforts to prevent transmission were unsuccessful), the time to transmission was increased 3-fold. A logical extension of these data is that expanded serological testing for evidence of HSV-2 infection, which strongly correlates with the presence of genital herpes, coupled with counseling regarding the potential for asymptomatic transmission of infection to uninfected partners and regarding the evidence that disclosure of one's infection status to partners will together reduce the likelihood of transmission, should now be encouraged as part of efforts to control this widespread infection.
We believe that currently available tools (serological testing, condom promotion, long-term suppressive antiviral therapy, and disclosure of infection status to sex partners) provide the means for developing a comprehensive national program to control the spread of genital herpes. There is certainly still additional operational research needed to understand how to best utilize currently available measures for control and logistical elements to identify and overcome. Nonetheless, as a first step, it is time for clear and specific serological screening recommendations for genital herpes. These recommendations must be accompanied by recommendations for counseling as well as funds to support the screening of populations most at risk for neonatal herpes and HIV infection-2 adverse outcomes associated with herpes. Knowledge is power, and knowledge can reduce the transmission risk of genital herpes. The time has come to translate knowledge of how to prevent genital herpes transmission into practice.
Knowledge of Partners' Genital Herpes Protects against Herpes Simplex Virus Type 2 Acquisition
The Journal of Infectious Diseases July 1, 2006;194:42-52
Anna Wald,1,2,3,5 Elizabeth Krantz,2 Stacy Selke,2 Ellen Lairson,2 Rhoda Ashley Morrow,2 and Judy Zeh4
Departments of 1Medicine, 2Laboratory Medicine, 3Epidemiology, and 4Statistics, University of Washington, and 5Program in Infectious Diseases, Fred Hutchinson Cancer Research Center, Seattle, Washington
(See the editorial commentary by Hook and Leone, above.)
ABSTRACT
Background. Prospective studies of herpes simplex virus type 2 (HSV-2) infection in discordant couples have shown a low rate of transmission. However, unlike partners with genital herpes in prospectively monitored couples, most persons who transmit genital herpes are not aware of having the infection.
Methods. Because HSV has a short incubation period and most persons who acquire genital herpes can identify the transmitting partner, a time-to-event design was used to assess risks of HSV acquisition among patients with newly acquired genital herpes.
Results. Among 199 persons with laboratory-documented newly acquired genital herpes, the median duration of the sexual relationship with the transmitting partner was 3.5 months, and the median number of sex acts before transmission was 40. The median time to HSV-2 acquisition was greater among participants whose partners disclosed that they had genital herpes, compared with participants whose partners did not disclose their status (270 vs. 60 days; P = .03). In multivariate models, having a partner who disclosed that he or she had genital herpes remained a strong protective factor against genital HSV-2 acquisition (hazard ratio, 0.48 [95% confidence interval, 0.25-0.91]).
Conclusion. These findings suggest that testing persons with HSV type-specific serologic assays and encouraging disclosure may result in a decreased risk of HSV-2 transmission to sex partners.
Although the prevalence of herpes simplex virus type 2 (HSV-2) infection has increased in the past 2 decades [1, 2], little is known about the dynamics of HSV-1 and HSV-2 transmission. Retrospective studies that evaluated source partners of persons with a first episode of genital herpes have suggested a short time between the initiation of sexual activity and infection [3]. In contrast, among stable, monogamous couples, the risk of transmission appears to be low [4, 5]. To provide a potential explanation for these apparently disparate data, we hypothesized that persons who know their HSV-2 status and disclose it to their sex partners are less likely to transmit HSV-2 than are persons who do not know their HSV-2 status. To address this question, we enrolled persons with documented newly acquired genital herpes and assessed the characteristics of the relationship that resulted in transmission. Because HSV has a short incubation period (2-12 days), most persons with newly acquired HSV infection can accurately identify the person who transmitted the infection to them [6, 7].
DISCUSSION
Our study of newly acquired genital herpes revealed several novel insights into the sexual transmission of HSV. Among persons with newly acquired genital herpes, the median duration of relationships was 3.5 months, which suggests that HSV is transmitted quickly in new relationships. Most persons who transmitted HSV-2 infection did not know that they had genital herpes; lack of awareness of the infection is consistent with published data from HSV-2 serosurveys [1]. Importantly, the risk of HSV-2 transmission was approximately halved when the source partner knew that he or she had genital herpes and informed the participant. These observations suggest that identifying persons with unrecognized HSV-2 infection may result in a decreased risk of transmission from such persons to new sex partners. The change in sexual behavior that accompanies disclosure and protects against STD transmission is not clear; however, the disclosure of HIV serostatus is considered to be important in the prevention of HIV transmission, and studies have clearly documented the complex issues involved in whether disclosure occurs [16]. Finally, our results show that HSV transmission occurs most often among partners who are monogamous and who consider the relationship to be steady. As such, messages about STD prevention that focus on decreasing the number of partners and avoiding concurrency are unlikely to affect HSV acquisition.
Our study used a design that has become the standard approach in studies of factors influencing time to pregnancy [15, 17-22]. We chose to censor participants at 1 year, to allow for the possibility that the source partner had acquired HSV from concurrent partners after the initiation of the relationship of interest. In that case, the duration of the relationship would not equal the duration of risk for HSV acquisition. An advantage of this approach is that we were able to study people who were in new relationships and those who may not have known that they or their partner had genital herpes. Such persons cannot be studied prospectively, because it would not be ethical to test people for HSV-2 and withhold test results. However, the disadvantage of the design is that, because we did not have access to a population of couples who did not transmit HSV, we were unable to calculate incidence rates, which is similar to studies of time-to-pregnancy that have yielded relative risks but not incidence rates [15].
Another corollary of our study design is that only persons with symptomatic genital herpes were enrolled, because we relied on clinical presentation for the initial diagnosis. A minority of newly acquired genital HSV-2 and HSV-1 infections result in a diagnosis at the time of acquisition, and the factors that distinguish between persons with symptomatic and asymptomatic HSV acquisition are poorly understood [23]. However, the enrollment of persons identified as seroconverting to HSV without clinical evidence of infection would necessitate prospective monitoring, obviating the possibility of this study design.
Gender and condom use, both of which have been recognized as determinants of HSV-2 transmission in prospective studies [5, 24, 25], did not influence the risk of transmission in this cohort. However, the increase in risk for women, compared with men, has varied widely among studies. For example, in 2 parallel studies of the subunit candidate vaccine, the risk of HSV-2 acquisition was 6-fold higher for women than for men who had a partner with genital herpes but was only 1.2-fold higher for women than for men among persons enrolled from STD clinics [5, 26]. Similarly, although condoms were shown to be protective in the candidate vaccine study, they were only marginally protective in a trial of valacyclovir for the prevention of HSV-2 transmission [4, 5]. Consistent condom use was unusual among our participants, which potentially suggests that those persons who use condoms regularly are protected against HSV-2 acquisition and, thus, could not become participants. Recent studies have emphasized the methodologic difficulties involved in the assessment of condom efficacy, and these may explain the variability in study findings [27, 28].
Despite the difficulty in relying on self-reported information about sexual behavior, we collected information about a relatively short period of time, during which recall of events should have been accurate and representative of the baseline status. Because the patients did not have an ongoing relationship with the clinician and were often distressed by their new diagnosis of genital herpes, social desirability bias was less likely to occur than in prospective studies that collect information on sexual activity and condom use after counseling about such activities.
The decrease in risk of transmission was noted in a previous study of HSV-2-discordant partners. In the 18-month trial of a candidate vaccine, the absolute risk of HSV-2 transmission decreased during follow-up, from 8.5 cases/100 person-years during the initial 150-day interval to 0.9 cases/100 person-years during the last 150-day interval [5]. Similarly, a relationship lasting >2.5 years before study entry and a duration of genital herpes lasting >2 years in the source partner were associated with a decreased risk of HSV-2 transmission in the valacyclovir clinical trial (although the former did not reach statistical significance) [4]. Of note, the median duration of relationships before study enrollment in these 2 trials was 18 and 24 months, respectively. In addition, the main reason for a lack of eligibility in the discordant-couple studies was the discovery that both partners were already infected with HSV-2, despite the clinical history of genital herpes in only 1 partner [4, 24, 29]. This serologic concordance at study entry suggests that transmission has already occurred, although it is also possible that the "susceptible" partner had acquired HSV-2 in a prior relationship. The present study demonstrates that most HSV-2 transmission events occur early in partnerships; thus, these couples are unlikely to participate in prospective studies.
That our study included fewer persons with genital HSV-1 infection limited its power to detect significant risks for HSV-1 transmission. However, the duration of the relationship before transmission was also short, and our results suggested that disclosing the presence of genital HSV-1 infection to partners may be protective. Of note, the source of newly acquired genital HSV-1 infections is usually unknown. Although, in some studies, a history of oral-genital contact appeared to be common among persons who acquired genital HSV-1 infection, studies of HSV transmission from women to neonates at delivery have shown that HSV-1 reactivation may be associated with a greater risk of transmission than HSV-2 reactivation [30-32]. As such, further studies are needed to define the risk of transmission from oral versus genital HSV-1 infection.
The time-to-event study design can yield a different type of insight than a prospective study, and, to our knowledge, it has not been used in the research of sexually transmitted pathogens or other infections. However, both a decrease in the rate of HIV transmission with the duration of relationship and a low frequency of HIV acquisition among persons in prospectively monitored HIV-discordant couples have been noted [33-37]. The low rate of observed transmission events results in difficulty in obtaining a firm understanding of important cofactors that may influence transmission. Couples who participate in such prospective studies may be fundamentally different from those who do not participate-they may be more committed to each other and more interested in preventing transmission, whether of HIV or HSV-2. Host factors that influence acquisition are also likely to vary, and individuals with the highest susceptibility may already have acquired infection by the time when they are identified as potential study participants. Finally, the course of infection in the source partner may result in varying levels of infectiousness over time. For HSV-2, early disease is associated with high shedding rates, and this may contribute to higher rates of transmission in new relationships, which then wane as the reactivation frequency declines [38, 39].
Can transmission from HSV-2-seropositive persons be prevented by serologic testing only? Probably not, but serologic testing of persons with clinically silent infection, as well as those at risk, is likely to be the cornerstone of any prevention strategy for HSV-2 infection. Accurate type-specific HSV serologic tests are commercially available, although their use is still limited [40]. Prospective studies of persons with HSV-2 antibodies have demonstrated that they have clinically and virologically active infection: most recognize recurrences and shed virus in the genital area [41-43]. More importantly, the persons with unrecognized HSV-2 infection are the source of most new infections, and our results suggest that transmission may be delayed by awareness of the infection. There is an ongoing debate about whether a wider use of HSV serologic tests in the general population is appropriate [40, 44, 45]. Concerns range from issues surrounding the accuracy of the tests, to the burden on health-care practitioners of providing appropriate counseling, to the psychosocial distress that may accompany a new diagnosis of genital herpes [46, 47]. Recent studies addressing these concerns have been reassuring: no lasting distress was identified in persons receiving a serologic diagnosis of HSV-2 infection [48-50]. Physicians should encourage patients to disclose their genital herpes to partners before the initiation of sexual activity, as currently recommended [8].
RESULTS
Of 536 patients referred for the study, 261 (49%) had documented newly acquired genital HSV-1 or HSV-2 infection. New acquisition of genital herpes was documented by detection of HSV from genital swabs in the absence of antibody to that type of virus in 97 patients (37%), by detection of HSV and seroconversion in 104 patients (40%), and by detection of seroconversion to HSV-1 or HSV-2 in the setting of clinically compatible lesions in 24 patients (9%). Thirty-six patients (14%) with newly acquired genital herpes were excluded because they did not provide data about sex partners. Among the remaining 275 patients, 180 had a first recognized recurrence of genital herpes, as documented by the presence of fully developed antibodies on the Western blot at first evaluation; 43 had alternative diagnoses; and 52 provided information insufficient to document the acquisition of infection (figure 1).
Among 225 participants with newly acquired genital HSV-1 or HSV-2 infection, 183 (81%) reported only 1 partner, 30 (13%) reported >2 partners, and 12 (5%) reported >3 partners in the 4 weeks preceding presentation. In 16 participants reporting 2 partners, laboratory evidence of concordant HSV infection in only 1 partner allowed for the identification of the transmitting partner. Thus, the analyses were conducted for 65 participants with genital HSV-1 infection and 134 participants with HSV-2 infection. Among these 199 participants with newly acquired HSV-1 or HSV-2 genital herpes, 127 (64%) were women and 72 (36%) were men. The median age was 26 years; 74% were white, 7% were African American, and 19% were of another race/ethnicity (table 1).
Partnership characteristics. Most persons who acquired genital herpes were in relatively new relationships, with 7% reporting a relationship lasting <1 week, 16% reporting a relationship lasting 1 week-1 month, and 64% reporting a relationship lasting 1-6 months before HSV acquisition. Only 42 participants (21%) reported relationships lasting >1 year and, thus, were censored at 1 year. The median duration of sexual relationships was 3.5 months (interquartile range [IQR], 1.5-10 months), and the median number of sex acts before HSV transmission was 40 (IQR, 15-75).
Despite the relatively short duration of the relationships, 166 participants (83%) characterized their relationship as steady (table 2). Condom use was infrequent, with 50% reporting condom use the first time they had sexual intercourse with the transmitting partner and 20% reporting condom use the last time they had sexual intercourse. Only 6% of persons with newly acquired genital HSV-1 infection and only 13% of persons with newly acquired HSV-2 infection reported that the partner disclosed to them that he or she had genital herpes.
The median time to HSV-2 acquisition was greater in participants whose partners informed them that they had genital herpes than in those whose partners did not (270 vs. 60 days, respectively; P = .03, log-rank test). By contrast, the median time to HSV-1 acquisition was 150 days for both groups (P = .41, log-rank test) (figure 2). Among the 22 partners who disclosed that they had genital herpes, 15 partners did so before having sex for the first time and 7 only did so after the initiation of sexual activity. Most participants with newly acquired genital herpes believed that the partner was truthful-only 9% of participants who acquired HSV thought that the partner knew that he or she had genital herpes but did not reveal it.
To confirm the identity of transmitting couples, we obtained HSV-2 isolates from 13 pairs of partners. Restriction enzyme analysis revealed that the pairs were indistinguishable in 9 cases, differed by 1 restriction enzyme in 1 case, and differed by 2 restriction enzymes in 3 cases. However, among the pairs with variations, none differed by >2 bands with any 1 enzyme, which indicates that they were within 1 genetic change of each other and, thus, were closely related.
Risk factors for HSV-2 acquisition. When we examined risk factors for HSV-2 acquisition, participants who were younger (hazard ratio [HR], 1.47 [95% confidence interval {CI}, 1.16-1.86] for each 10 years younger) and those who were younger than their partner (HR, 1.82 [95% CI, 1.10-3.02]) were at a significantly increased risk of HSV-2 acquisition (table 3). Participants with more education (HR, 0.46 [95% CI, 0.28-0.76] for those with some college vs. a high school degree or less; HR, 0.59 [95% CI, 0.37-0.94] for those with a 4-year college degree or more vs. those with a high school degree or less), participants who discussed STDs or genital herpes with their partner (HR, 0.51 [95% CI, 0.33-0.79] for discussing genital herpes vs. not), and those who were told by their partner that he or she had genital herpes (HR, 0.55 [95% CI, 0.31-0.98]) were at a significantly lower risk of HSV-2 acquisition. Sex, race, and sexual preference were not associated with the risk of HSV-2 acquisition. Condom use was not found to be protective in any analyses.
In multivariate models, having a partner who disclosed that he or she had genital herpes remained a strong protective factor against genital HSV-2 acquisition (HR, 0.48 [95% CI, 0.25-0.91]). In addition, persons with more education and those who were older remained at a relatively lower risk of HSV-2 acquisition (table 3). After adjustment for other characteristics, women were at a somewhat increased risk for HSV-2 acquisition (HR, 1.47 [95% CI, 0.97-2.23]).
In a sensitivity analysis that reclassified participants who reported that their partner disclosed having genital herpes after the first time they had sex as not having been informed, the protective effect of having been told that the partner had genital herpes was greater than in the original analysis (adjusted HR, 0.34 [95% CI, 0.15-0.81]). Thus, having been told before engaging in sex was associated with a lower risk of HSV-2 acquisition than having been told at any time during the relationship, although the CIs overlapped.
Risk factors for HSV-1 acquisition. Similar to the analysis that examined risk factors for HSV-2 acquisition, we explored the factors that may be important for HSV-1 acquisition among 65 participants with genital HSV-1 infection. In univariate analysis, younger age (HR, 2.22 [95% CI, 1.03-4.78] for <32 vs. >32 years) and homosexual or bisexual preference (HR, 3.15 [95% CI, 1.06-9.35]) were significantly associated with a higher risk of HSV-1 acquisition (table 4). Talking about STDs with their partner had a protective effect (HR, 0.46 [95% CI, 0.24-0.87]). Participants who were told that their partner had genital herpes had a decreased risk of HSV-1 acquisition, although this result was not significant (HR, 0.56 [95% CI, 0.14-2.33]). However, only 4 of 65 participants with HSV-1 infection reported that their partner said he or she had genital herpes. Other factors-including race, education, and condom use with the transmitting partner-were not significantly associated with risk of HSV-1 acquisition.
Multivariate analyses adjusting for sex, age, history of STDs, number of lifetime partners, and suspicion of partner engaging in concurrent relationships revealed a stronger point estimate for having been told that their partner had genital herpes, although this result was still not statistically significant (adjusted HR, 0.35 [95% CI, 0.07-1.70]). Younger age remained significantly associated with HSV-1 acquisition in the multivariate model (table 4). Women were at a lower risk of HSV-1 acquisition than were men (adjusted HR, 0.48 [95% CI, 0.24-0.97]); this was partially explained by the higher risk of HSV-1 acquisition among men who had sex with men. A history of STDs, number of lifetime partners, and suspicion of partners having concurrent relationships were not statistically significant; however, because they substantially influenced the estimate for being told that the partner had genital herpes, they were retained in the multivariate model.
PATIENTS, MATERIALS, AND METHODS
Study participants and procedures. Patients presenting to the University of Washington Virology Research Clinic or Sexually Transmitted Disease (STD) clinic in Seattle with a first episode of genital lesions consistent with herpes were invited to participate. The participants underwent clinical examination and collection of swabs (for HSV culture and polymerase chain reaction [PCR]) and blood (for detection of antibodies against HSV-1 and HSV-2). Using a standardized questionnaire, the participants were asked about the number of sex partners they had during the 4 weeks preceding the development of genital lesions and specific questions about relationships with up to the 3 most recent sex partners. Treatment with antiviral therapy and counseling were provided as recommended elsewhere [8], and blood was drawn at 3 months for the detection of seroconversion to HSV [9]. None of the participants reported being HIV positive, and most had tested HIV negative in the past. The University of Washington institutional review board approved the protocol, and all participants signed a consent form.
Newly acquired genital HSV infection was defined as clinically diagnosed genital herpes with laboratory documentation of new HSV-1 or HSV-2 infection. The laboratory documentation included detection of HSV in a person who lacked antibodies to the HSV type found in genital secretions or demonstration of seroconversion to HSV-1 or HSV-2. Inclusion or exclusion of partnerships from analysis was based on a definition of transmitting partners that was specified before analysis. Sex partners were defined as transmitting partners if they were the only sex partner during the 4 weeks preceding the development of genital herpes. Patients reporting >1 partner during that 4-week interval were included only if (1) all partners were evaluated and only 1 had virologic or serologic evidence of the infection that was transmitted or (2) all but 1 partner was evaluated and none of the partners who were evaluated had laboratory evidence of the infection that was transmitted. In that case, the single partner who was not evaluated was identified as the transmitting partner.
Laboratory methods. HSV Western blot was used to identify type-specific antibodies to HSV-1 and HSV-2 [10]. Seroconversion was identified as the acquisition of antibodies in paired serum samples. HSV culture, typing of the isolates, HSV DNA PCR, and restriction endonuclease analysis were performed as described elsewhere [11-14].
Statistical analyses. The acquisition of HSV-1 and HSV-2 was analyzed separately but with the same methodology. The outcome-the time to acquisition of genital herpes-was defined as the number of days from the beginning of the sexual relationship with the transmitting partner until acquisition [15]. Relationships lasting >1 year were censored at 1 year. The main predictor-whether the source told the participant that he or she had genital herpes-was modeled as telling versus not telling. Kaplan-Meier curves, log-rank tests, and proportional-hazards models were used to assess relationships between the main predictor, as well as other patient characteristics, and the time to acquisition of genital herpes. Because of the study design, the main predictor was ascertained at the time of the outcome. To allow for appropriate survival modeling, we assumed that this covariate information was indicative of the baseline status, defined at the beginning of the relationship. However, we also performed a sensitivity analysis for this variable in which we reclassified subjects by their true baseline status. Therefore, persons whose partners told them that they had genital herpes after the first time they had sex were then categorized in sensitivity analyses as not having been told.
|
|
| |
| |
|
|
|
